Tim Moynihan, Director of Product Marketing, Intel
What has been Intel's involvement with speech technologies?
Tim Moynihan Intel has been involved with speech technology and media processing for many years. For years Intel has enabled the processor to natively handle media streaming required for many video and audio applications including speech. Through Intel's investments in the communications market, we brought significant experience in communications technology into the company. Many of the senior executives in Intel's Communications Group, as well as the engineers working to push the limits of innovation in Intel communications labs around the world have spent decades developing and advancing speech technologies. The power of Intel's industry standard processors has now advanced to the point that computationally intensive speech applications (like speech recognition as well as real time audio media streaming) can now be reliably performed in scale at viable price/performance points. Our efforts are focused in three major areas: technology, ecosystem and standards. Our goal is to lower the cost and complexity of speech applications by creating a foundation of open, modular computing and communications building blocks on which these solutions will run. Our technology efforts are focused on key areas of speech solutions including telephony interface products and speech optimization on our two primary computing architectures: Intel Architecture for servers and PCA for clients. Intel Labs continues to conduct primary research in many areas related to speech. Intel continues to actively participate in industry forums and standards organizations such as the VoiceXML Forum, the SALT forum, W3C, and ITEF. We have recently launched the Intel Communication Alliance to support the community of developers in the communications market. Finally, Intel Capital continues to explore investment opportunities in key areas in the industry.
Discuss what the latest announcement concerning Intel NetMerge Call Manager and what it means for speech deployments.
TM We began publicly discussing our relationship with the Microsoft speech effort about a year ago. This week at SpeechTEK we are making our first, formal announcement of a product related to this effort: the Intel NetMerge Call Manager. Microsoft has the stated goal of making speech mainstream. With the Intel NetMerge Call Manager, Intel is focusing on simplifying the telephony integration component of the reference architecture. It is an off-the-shelf software component that provides a fast and easy way for developers to integrate Intel's leading telephony boards with the Microsoft Speech Platform. In terms of the Microsoft speech reference architecture, the Intel NetMerge Call Manager is a Telephony Interface Manager (TIM). The Intel NetMerge Call Manager fully complies with Microsoft's Telephony Application Server (TAS) specification. It offers developers a robust feature set as well as tools for configuration, administration, and diagnostics.
What is Intel hoping to accomplish with Intel NetStructure Host Media Processing?
TM Intel NetStructure Host Media Processing is a new way to deliver voice media processing capabilities. By utilizing software developed to fully utilize the media processing capabilities of the Intel Pentium and Celeron processors to replace specialty hardware (Digital Signal Processing or DSP chips), developers will be able to more rapidly develop applications, and VARs and end users be able to reduce the time and cost associated with installing, configuring and maintaining those solutions. The software business model and the flexibility of being able to use only the exact number of media resources required are the keys to these efficiencies.
Is Intel and Microsoft strengthening their existing partnership ties with these announcements? What does this mean for your existing partners?
TM Intel and Microsoft are two companies pursuing similar opportunities in markets based on a common belief that industry standards offer the opportunity to deliver high volume, low cost, reliable solutions. Speech is a market segment, like e-mail, or data base management, that we both believe can be best addressed using industry standard platforms. Since both of our companies are making complimentary building blocks for these solutions, our research and innovation can benefit an entire ecosystem of related companies pursuing opportunities in tool kits, applications development, Distribution, Value Added Reselling, and application hosting - to name a few. I would characterize this as a step forward in realizing our vision of simplified speech deployments. The Intel and Microsoft engineering and development teams have been closely collaborating on this effort for well over a year. Intel's product has been in field trials since late last year. Our existing customers should view this as a validation of our mutual commitment to the success of the program.
What markets are you targeting with these releases?
TM Primarily these product release announcements are aimed at informing the community of application developers that there is a new and faster way to build solutions. Since many developers are already developing and preparing solutions built on these products, we are also interested in reaching the other companies that will be required to sell, distribute, configure, install and maintain these solutions as the market develops. Developers should understand the benefits of these new products are either communications solutions developers, who have used specialty or purpose built hardware to deliver functionality in the past. Or Commercial Off The Shelf (COTS) application developers, who can now use their existing skills and capabilities to develop and deliver voice communications solutions.
During a recent panel and you were somewhat a lone voice concerning the short-term impact of speech in mobile applications. What makes you believe that speech technologies will be increasingly deployed in a mobile environment?
TM Speech, and speech recognition in particular, have been a hot topic in mobile telephony for over 10 years. Speech in enterprise mobile environments using 802.11 or other mobile infrastructures is a new hot topic today. Using speech to control calls helps to simplify the act of completing calls thus improving safety, and making the mobile phone easier to use and thus more indispensable. Using voice technologies to allow access to corporate information (e.g., auto attendant and IVR applications) can be said to provide ease of use benefits to both mobile and stationary users. The rapid installation of WiFi access points globally is evidence of the mobility infrastructure being built out to support applications designed for the mobile user. Speech based solutions will be a key part of the mobile solution providing navigation and location-based input and output to users.
What should the industry be doing now to ensure this adoption happens?
TM In a sense, the industry should continue doing what it's doing:
· Evangelizing the benefits of speech technology
· Aggressively supporting standards initiatives
· Focusing on delivering tools and applications, and services that simplify development and deployment Additionally, we all need to understand why standard servers are now capable of performing voice media processing tasks (thanks to Moore's law, and the expanded power and improved price performance of that power) and deliver that message to the market. Standard application development tools and skills can now be used to rapidly develop voice/speech-based applications that are reliable and cost effective.
Discuss your thoughts on what IT enterprise customers can expect from Intel and their speech efforts over the next 15 months.
TM Enterprises indirectly see the results of our efforts thought he great products and services introduced by our fellow travelers in the speech ecosystem. For one thing, they will see continued performance improvements and modularization for telephony-based speech solutions. Enterprises will also see an increased emphasis on mobility applications. This means client devices capable of supporting mulitmodal user interfaces combined with more ecosystem development activities focused on this compelling application segment. Enterprises should expect three things form the technology suppliers such as Intel: More innovation - better more efficient and powerful products
More standardization - Intel continues to work tirelessly on standards bodies W3C Voice XML forum, SALT Forum, PICMIG, IEEE, etc to ensure that standards are developed and ratified that allow the industry to advance its capabilities and deliver functionality to end users
More market support - Intel's Communications Alliance is being launched to be support our partners and suppliers in developing solutions for the communications market segment